Excitation of Seismic Waves by the Atmosphere: Monitoring Severe Weather with Modern Digital Seismic Data
When sufficiently strong, hurricanes and tornadoes generate significant observable ground motions through pressure changes at the surface. Under the proper circumstances, these signals are recorded by seismometers and can provide insight into the storm events, which we examine in this thesis. First, we used two dense seismic arrays, Earthscope’s Transportable Array (TA) and the Southern California Seismic Network (SCSN) to examine the case of landfallen hurricanes through the TA. Through the study of Hurricane Isaac, which made landfall through the TA in 2012, we can observe its decay as it weakened and dissipated over time in the seismic data within the frequency range of 0.01-0.02 Hz. With this data, we develop a stochastic theory for the generation of seismic waves by a hurricane.
In further examining the seismic and barometric signals generated by Hurricane Isaac and Tropical Storm Lee (2011), we also identify the existence of a threshold pressure. Above a certain level of atmospheric surface pressure, there is dependence in the seismic wave generation. While this and the theory of seismic wave generation are novel results, these observations do not hold many practical applications towards hurricane monitoring, which we go into further detail in the final potion of this thesis.
In regards to tornadoes, we identified seismic signals corresponding to an EF5 tornado that occurred on May 22, 2011 in Joplin, Missouri. This signal was recorded by the TA station that was closest to the tornado track. We modeled this seismic signal at low frequencies (below 0.1 Hz) by assuming an equivalent vertical force and found that the amplitude of the seismic signal corresponds to the reported intensity of the storm. Further analysis of tornadoes in this way could provide a quantitative method of measuring tornado strength using seismic data.
Finally, using the SCSN, we performed a backprojection of 0.2 Hz P-waves of hurricanes over the ocean between 2011-2017, in order to test the possibility of tracking them using seismic data. We find that for many strong hurricanes, the backprojection results in a P-wave amplitude peak that can be associated with the storm. There is, however, a dependence on the size of the storm, as backprojection tracking performs best on the largest hurricanes that were examined in this study.